Decision Guide BPO
The cheapest rate card is not always the cheapest contract. Compare pricing units, define what counts as resolved, check what is included, and model your own workload.
The short answer: customer service outsourcing cost depends heavily on the pricing model. Dedicated per-agent pricing usually suits steady, predictable volumes where you want a consistent team that knows your product. Per-hour pricing suits workloads that vary by season or day, provided the provider can flex staffing. Per-contact and per-resolution pricing suit well-defined, repeatable contacts where you want to pay for output and the provider can control efficiency. Many contracts combine models, for example a dedicated core team with per-hour or per-contact overflow for peaks. The total customer support outsourcing cost depends as much on what is included, and on how "resolved" is defined, as on the headline rate.
This guide defines the pricing units, compares dedicated, shared and outcome-based models, lists the costs that sit outside the rate, and models three workloads with explicit assumptions.
Per agent (dedicated). You pay a monthly fee for each full-time agent assigned to your account. The agents work only on your customers, learn your product deeply and follow your processes. You pay whether contact volume is high or low.
Per hour. You pay for agent hours, either dedicated or shared with other clients. Check whether billing covers logged-in hours or only productive hours, and how breaks, training and meetings are treated.
Per contact or per ticket. You pay a fixed price for each contact handled, sometimes varying by channel. Some providers bill per minute for phone. The provider absorbs inefficiency, so prices typically include a margin for that risk.
Per resolution or per outcome. You pay only when an issue is resolved according to an agreed definition. This model is common for AI agents and is appearing in human support for well-defined contact types.
Hybrid and performance-linked. A base fee plus bonuses or penalties tied to service levels, quality scores or customer satisfaction.
If any part of the price depends on resolution, the definition is the contract. Agree in writing:
A loose definition rewards closing tickets rather than solving problems. As a dated example of outcome pricing in AI support, Intercom's pricing page listed its Fin AI agent at $0.99 per outcome on October 27, 2026; the page itself does not define the outcome, which illustrates why definitions need checking in the contract.
| Factor | Dedicated per agent | Shared per hour or per contact | Per resolution or outcome |
|---|---|---|---|
| Cost predictability | High: fixed monthly cost | Moderate: varies with volume | Varies with resolved volume |
| Flexibility for peaks | Low unless contract allows ramping | High if provider has capacity | High |
| Product knowledge | Highest | Lower; agents serve several clients | Depends on provider model |
| Quality control | Direct; you know the team | Harder; shared agents | Strong incentive to resolve, but risk of gaming definitions |
| Best for | Steady volume, complex products, brand-sensitive service | Variable volume, simpler contacts | Well-defined contact types |
| Watch for | Paying for idle capacity | Inconsistent quality, rate creep on overtime | Narrow definitions, cherry-picking easy contacts |
| Pricing unit | Dedicated agent | Hour / contact | Resolution |
|---|---|---|---|
| What is billed | Committed team capacity | Staffed time or contact count | A defined resolved issue |
| Workload fit | Steady, product-specific work | Variable or repeatable demand | Repeatable, controllable outcomes |
| Knowledge | A consistent named team | Depends on shared team setup | Depends on delivery model |
| Main trade-off | Idle capacity at low demand | Availability and quality at peaks | Reopens, exclusions and definitions |
| Scope factors | Training, tools, QA, management | Training, tools, QA, management | Training, tools, QA, management |
Two quotes with similar rates can differ widely in total cost. Ask what each includes:
The scenarios below use invented rates purely to show how the models behave. They are not market quotes or Growix pricing. Rates vary widely by location, language, channel, complexity and contract length. For a US reference point, the Bureau of Labor Statistics reports a median wage of $21.53 an hour for customer service representatives in May 2025; provider rates add benefits, management, facilities, tools and margin, and differ substantially by delivery location, as our guide to offshore vs nearshore vs onshore outsourcing explains.
Assumed rates for all scenarios: dedicated agent $3,200 a month; shared hours $20 an hour; per contact $4.50; per resolution $7.00; each dedicated agent provides about 160 productive hours a month.
6,000 contacts a month, 7 minutes average handling time, 75% occupancy.
With steady volume and efficient handling, paying for capacity is cheapest. Output-based models cost more because the provider prices in the risk of variation.
3,000 contacts a month for ten months, 9,000 a month in November and December, so 48,000 a year. Same handling time and occupancy.
Seasonality makes flexibility valuable. A dedicated team sized for the peak wastes most of its cost; a core team with seasonal ramp or flexible hours is far cheaper, provided the provider can actually supply trained people for the peak.
800 tickets a month, 25 minutes average handling time, 70% occupancy.
For complex, low-volume work, the price difference between models is small; product knowledge and escalation quality matter far more. A dedicated team usually builds that knowledge faster. Our guide to SaaS technical support outsourcing covers the handoff design.
Real contracts rarely use one model for everything. Common blends include:
Blends work best when each component has its own scope and reporting, so you can see what each part of the spend delivers.
The first-year price is only part of the story. Check:
Many providers offer a pilot period. Use it to test the assumptions that drive cost: actual handling times on your contacts, quality scores against your scorecard, escalation accuracy and the provider's ability to staff to forecast. Agree the pilot's measures and decision date in advance, and compare results with your own team's baseline on the same contact types. A pilot that only proves agents can log in and answer easy questions tells you little about the full contract.
When Growix may not be the right fit: if you need fewer than one full-time agent's worth of coverage, or only occasional after-hours cover for a handful of contacts, a shared-service or answering-service provider may be more economical than a dedicated team.
It depends on the pricing model, location, skills, hours of coverage and your contact volumes. Providers price per dedicated agent, per hour, per contact or per resolution, and two quotes with similar rates can differ widely once training, tools, quality assurance and management are included. Model each quote against your own workload.
Per-agent pricing usually suits steady volumes and a consistent team, per-hour pricing suits seasonal or variable workloads, and per-contact or per-resolution pricing suits well-defined, repeatable contacts with a tight definition of resolved. Many contracts blend a dedicated core with flexible overflow.
Check onboarding and ongoing training, team leads and quality analysts, workforce planning, reporting, tools and licenses, after-hours premiums, minimum commitments, annual rate increases and notice periods. Ask how each is priced, not just the headline agent rate.
Export three to twelve months of contact data and calculate volume, handling time and seasonality by contact type. With that, any provider can quote accurately, and you can compare models on total cost rather than headline rates. If AI is part of your plan, see AI customer support vs human agents vs a hybrid team for how the split affects human workload and cost.
Growix's Customer Support Outsourcing service delivers email, chat and phone support to your tone, policies and service levels, and our Technical Support Outsourcing team provides Level 1 and Level 2 technical support with escalation to your engineers. If missed calls and inconsistent service differ by location, see customer journey mapping for multi-location businesses.
The useful comparison is cost per resolved issue at an acceptable quality, not cost per contact. Segment requests by complexity and risk, then design the AI-to-human handoff.
Outsourced support succeeds or fails at the handoffs. Define what each tier owns, share access and knowledge safely, set escalation and incident rules, then audit quality and engineering load.
Location changes more than the hourly rate. Compare loaded cost, working-hour overlap, language and culture fit, data handling and supervision, then score providers by work type.
